Quote from: jedi12345 on Today at 09:06:36 PM

Quote from: madogss on Today at 08:30:31 PM

Quote from: jedi12345 on Today at 07:51:01 PM

Hey Puzzlers

What would be the best offline way to test a big list of either Decimal or Binary possibilities I have.

Just CPU testing, I could buy a GPU if its like 100x faster than CPU.

Happy to give the 1-3 people 0.1btc that help me when I solve puzzle 69. (If that's allowed on this forum)

Note offline solving only, no pools or online, happy to download any requirements first. - This pc will not be connected to the internet once ready to solve the problem.

Cheers

For cpu you want to use keyhunt or keyhunt-cuda. here's the links https://github.com/WanderingPhilosopher/KeyHuntCudaClient https://github.com/albertobsd/keyhunt

You can also find pre compiled ones for keyhunt on github.

cpu averages about 3.5 million keys a second per thread if using above software or I believe according to ktimesg 10 million keys a second per thread with custom software.

Gpu's are way faster a 3080 does about 2200MK/s or 2.2 billion keys a second.

Best way to test a big list is to write a python script to read the doc and import the possibilities into a list then in a loop convert the decimal or binary into hexadecimal get rid of the 0x in the front and run keyhunt or kehunt-cuda in that range using a subprocess.

Ok few questions before I get a GPU, because I will have around 200 trillion combinations maybe more.

Is the 4090 or 5090 way faster or not worth the $$ or is the 3090 just the best bang for your buck does it have to be Nvidia as AMD have good price to performance.

Converting into a Hexadecimal without Ox sounds easy enough. Hopefully python can output 100 trillion numbers without issues.

Could you give me a basic run though and how to set up and use keyhunt-cuda on a premade hexadecimal file (using a subprocess). Would it be txt file or a py file or a json file.

Also I take it this computation can be done offline once setup?

Cheers

Python is slow. C++ is 100 times faster than python. All computation can be done offline. I think RTX 5070 is the best best bang for the buck. It is as fast as 4090 and almost one fourth the price.